(PS8-98) Transdiagnostic Psychopathology Networks Among Patients in a Digital Mental Health Intervention Selection Trial

Digital interventions have been shown to be effective in treating anxiety and depression (Huang et al., 2024). However, optimal treatment selection and delivery may benefit from a personalized approach that considers the interconnectivity of symptoms across psychiatric disorders (Mohr et al., 2018). Notably, dimensional and transdiagnostic models of psychopathology point to the interrelated nature of symptoms across disorders and the importance of identifying and targeting core, shared features (Barlow et al., 2021; Kotov et al., 2017). This study utilizes network analysis to explore how symptoms of anxiety (measured by the GAD-7), depression (measured by the PHQ-9), and well-being (measured by the WHO-5) are interrelated before treatment begins. Data are drawn from a clinical trial on digital treatment selection optimization among patients (N=288) with emotional disorders within the Kaiser Permanente health symptom network (NCT05567640). The analysis identifies central symptoms and bridge nodes (connecting different symptom clusters) that have the greatest influence within the symptom network (Borsboom, 2017). Preliminary findings identify key symptoms that significantly impact the network and highlight bridge symptoms that connect different clusters of distress. The findings indicate that cognitive-affective symptoms, such as negative self-perception and excessive worry, play a central role in sustaining emotional distress more broadly. There are strong connections between symptoms of anxiety and depression, highlighting their overlap and supporting a transdiagnostic approach to mental health treatment. Bootstrap analyses indicated that the symptom relationships and centrality estimates were stable, suggesting that the identified highly influential symptoms are robust treatment targets. These insights could improve the personalized allocation of digital interventions through precision psychiatry models (Kessler & Luedtke, 2021). Additionally, because engagement is critical for adherence to digital interventions (Baumel et al., 2019), tailoring treatments based on the initial symptom network structure may help enhance patient engagement and improve clinical outcomes. While cross-sectional network analysis provides valuable insights into psychopathology and informs treatment decisions, it cannot distinguish between symptom fluctuations within individuals and differences between individuals. Future research should employ longitudinal modeling to capture these dynamic processes, enhancing precision in treatment selection and optimization.